Bayesian Nonparametric Causal Inference: Information Rates and Learning Algorithms
We investigate the problem of estimating the causal effect of a treatment on individual subjects from observational data; this is a central problem in various application domains, including healthcare, social sciences, and online advertising. Within the Neyman-Rubin potential outcomes model, we use...
Uloženo v:
| Vydáno v: | IEEE journal of selected topics in signal processing Ročník 12; číslo 5; s. 1031 - 1046 |
|---|---|
| Hlavní autoři: | , |
| Médium: | Journal Article |
| Jazyk: | angličtina |
| Vydáno: |
New York
IEEE
01.10.2018
The Institute of Electrical and Electronics Engineers, Inc. (IEEE) |
| Témata: | |
| ISSN: | 1932-4553, 1941-0484 |
| On-line přístup: | Získat plný text |
| Tagy: |
Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
|
| Abstract | We investigate the problem of estimating the causal effect of a treatment on individual subjects from observational data; this is a central problem in various application domains, including healthcare, social sciences, and online advertising. Within the Neyman-Rubin potential outcomes model, we use the Kullback-Leibler (KL) divergence between the estimated and true distributions as a measure of accuracy of the estimate, and we define the information rate of the Bayesian causal inference procedure as the (asymptotic equivalence class of the) expected value of the KL divergence between the estimated and true distributions as a function of the number of samples. Using Fano's method, we establish a fundamental limit on the information rate that can be achieved by any Bayesian estimator, and show that this fundamental limit is independent of the selection bias in the observational data. We characterize the Bayesian priors on the potential (factual and counterfactual) outcomes that achieve the optimal information rate. We go on to propose a prior adaptation procedure (which we call the information-based empirical Bayes procedure) that optimizes the Bayesian prior by maximizing an information-theoretic criterion on the recovered causal effects rather than maximizing the marginal likelihood of the observed (factual) data. Building on our analysis, we construct an information-optimal Bayesian causal inference algorithm. This algorithm embeds the potential outcomes in a vector-valued reproducing Kernel Hilbert space, and uses a multitask Gaussian process prior over that space to infer the individualized causal effects. We show that for such a prior, the proposed information-based empirical Bayes method adapts the smoothness of the multitask Gaussian process to the true smoothness of the causal effect function by balancing a tradeoff between the factual bias and the counterfactual variance. We conduct experiments on a well-known real-world dataset and show that our model significantly outperforms the state-of-the-art causal inference models. |
|---|---|
| AbstractList | We investigate the problem of estimating the causal effect of a treatment on individual subjects from observational data; this is a central problem in various application domains, including healthcare, social sciences, and online advertising. Within the Neyman-Rubin potential outcomes model, we use the Kullback-Leibler (KL) divergence between the estimated and true distributions as a measure of accuracy of the estimate, and we define the information rate of the Bayesian causal inference procedure as the (asymptotic equivalence class of the) expected value of the KL divergence between the estimated and true distributions as a function of the number of samples. Using Fano's method, we establish a fundamental limit on the information rate that can be achieved by any Bayesian estimator, and show that this fundamental limit is independent of the selection bias in the observational data. We characterize the Bayesian priors on the potential (factual and counterfactual) outcomes that achieve the optimal information rate. We go on to propose a prior adaptation procedure (which we call the information-based empirical Bayes procedure) that optimizes the Bayesian prior by maximizing an information-theoretic criterion on the recovered causal effects rather than maximizing the marginal likelihood of the observed (factual) data. Building on our analysis, we construct an information-optimal Bayesian causal inference algorithm. This algorithm embeds the potential outcomes in a vector-valued reproducing Kernel Hilbert space, and uses a multitask Gaussian process prior over that space to infer the individualized causal effects. We show that for such a prior, the proposed information-based empirical Bayes method adapts the smoothness of the multitask Gaussian process to the true smoothness of the causal effect function by balancing a tradeoff between the factual bias and the counterfactual variance. We conduct experiments on a well-known real-world dataset and show that our model significantly outperforms the state-of-the-art causal inference models. |
| Author | Alaa, Ahmed M. van der Schaar, Mihaela |
| Author_xml | – sequence: 1 givenname: Ahmed M. orcidid: 0000-0001-9936-7141 surname: Alaa fullname: Alaa, Ahmed M. email: ahmedmalaa@ucla.edu organization: Department of Electrical Engineering, University of California, Los Angeles, Los Angeles, CA, USA – sequence: 2 givenname: Mihaela surname: van der Schaar fullname: van der Schaar, Mihaela email: mihaela.vanderschaar@eng.ox.ac.uk organization: Department of Engineering Science, University of Oxford, Oxford, U.K |
| BookMark | eNp9kE1PwkAQhjcGEwH9A3pp4rm4H9126w2JHxiiBvDczLZTXEK3uFsO_HtbIB48eJr3MM87mWdAera2SMg1oyPGaHr3ulguPkacMjXiKlJc0DPSZ2nEQhqpqNdlwcNISnFBBt6vKZVJzKI-mT_AHr0BG7zVdgsOKmycyYMJ7Dxsgqkt0aHN8b6LtaugMbUN5tCgD8AWwQzBWWNXwXizqp1pvip_Sc5L2Hi8Os0h-Xx6XE5ewtn783QynoU5T2UTYkLLnCkUKSujWIMSSV4ornSpEs6gjPOCayi0BJpqigXEKaMCJGqNWupUDMntsXfr6u8d-iZb1ztn25MZZ0y2fZLKdksdt3JXe--wzHLTHL5oHJhNxmjWGcwOBrPOYHYy2KL8D7p1pgK3_x-6OUIGEX8BJVSiIil-AFP2gCg |
| CODEN | IJSTGY |
| CitedBy_id | crossref_primary_10_1002_int_22573 crossref_primary_10_1093_aje_kwab047 crossref_primary_10_1016_j_eswa_2022_117116 crossref_primary_10_1109_TIM_2024_3488141 crossref_primary_10_1057_s41270_023_00237_3 crossref_primary_10_1109_ACCESS_2019_2932390 crossref_primary_10_3390_electronics13214236 crossref_primary_10_1109_COMST_2019_2943405 crossref_primary_10_1002_cpt_1907 crossref_primary_10_1109_TSP_2025_3585825 |
| Cites_doi | 10.1073/pnas.1510489113 10.1002/sim.6265 10.1109/18.54897 10.1002/sim.4322 10.1073/pnas.1303102110 10.1109/TIT.2017.2773474 10.1214/08-EJS273 10.1006/jmva.1997.1659 10.1214/15-AOS1341 10.1080/01621459.1984.10478078 10.1198/jcgs.2010.08162 10.1214/074921708000000156 10.1109/TIT.2007.915707 10.1214/aos/1018031206 10.1561/2200000036 10.1214/aos/1017939142 10.1090/S0002-9939-1966-0193624-0 10.2202/1557-4679.1308 10.1214/aos/1176345969 10.1007/BF00050848 10.1257/aer.p20171038 10.2139/ssrn.3048177 10.1037/h0037350 10.1007/s11749-007-0075-5 10.1214/009053607000000613 10.1214/009053604000000300 10.1214/11-AOS920 10.1214/15-EJS1078 10.1007/978-94-017-2973-4_7 10.1002/sim.7623 10.3982/ECTA11293 10.1080/10485252.2017.1339309 10.1214/aos/1176345206 10.1177/0081175012452652 10.1080/10618600.2017.1356325 10.1070/SM1967v002n03ABEH002343 10.1007/3-540-44581-1_27 |
| ContentType | Journal Article |
| Copyright | Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2018 |
| Copyright_xml | – notice: Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2018 |
| DBID | 97E RIA RIE AAYXX CITATION 7SP 8FD H8D L7M |
| DOI | 10.1109/JSTSP.2018.2848230 |
| DatabaseName | IEEE All-Society Periodicals Package (ASPP) 2005–Present IEEE All-Society Periodicals Package (ASPP) 1998–Present IEEE Xplore CrossRef Electronics & Communications Abstracts Technology Research Database Aerospace Database Advanced Technologies Database with Aerospace |
| DatabaseTitle | CrossRef Aerospace Database Technology Research Database Advanced Technologies Database with Aerospace Electronics & Communications Abstracts |
| DatabaseTitleList | Aerospace Database |
| Database_xml | – sequence: 1 dbid: RIE name: IEEE Xplore url: https://ieeexplore.ieee.org/ sourceTypes: Publisher |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Engineering |
| EISSN | 1941-0484 |
| EndPage | 1046 |
| ExternalDocumentID | 10_1109_JSTSP_2018_2848230 8387845 |
| Genre | orig-research |
| GrantInformation_xml | – fundername: NSF grantid: ECCS1462245; ECCS1533983; ECCS1407712 |
| GroupedDBID | -~X 0R~ 29I 4.4 5GY 5VS 6IK 97E AAJGR AARMG AASAJ AAWTH ABAZT ABQJQ ABVLG ACIWK AENEX AETIX AGQYO AGSQL AHBIQ AKJIK AKQYR ALMA_UNASSIGNED_HOLDINGS ATWAV BEFXN BFFAM BGNUA BKEBE BPEOZ CS3 DU5 EBS EJD F5P HZ~ IFIPE IPLJI JAVBF LAI M43 O9- OCL RIA RIE RNS AAYXX CITATION 7SP 8FD H8D L7M RIG |
| ID | FETCH-LOGICAL-c295t-e70fc18e391f46ba837cd828bf8721af6cd2badb5a09b0eda69103a5ebbeb5b93 |
| IEDL.DBID | RIE |
| ISICitedReferencesCount | 23 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000446341300016&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 1932-4553 |
| IngestDate | Mon Jun 30 10:18:47 EDT 2025 Tue Nov 18 21:07:00 EST 2025 Sat Nov 29 04:10:32 EST 2025 Wed Aug 27 02:36:30 EDT 2025 |
| IsDoiOpenAccess | false |
| IsOpenAccess | true |
| IsPeerReviewed | true |
| IsScholarly | true |
| Issue | 5 |
| Language | English |
| License | https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html https://doi.org/10.15223/policy-029 https://doi.org/10.15223/policy-037 |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-c295t-e70fc18e391f46ba837cd828bf8721af6cd2badb5a09b0eda69103a5ebbeb5b93 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14 |
| ORCID | 0000-0001-9936-7141 |
| PQID | 2115837505 |
| PQPubID | 75721 |
| PageCount | 16 |
| ParticipantIDs | ieee_primary_8387845 proquest_journals_2115837505 crossref_citationtrail_10_1109_JSTSP_2018_2848230 crossref_primary_10_1109_JSTSP_2018_2848230 |
| PublicationCentury | 2000 |
| PublicationDate | 2018-10-01 |
| PublicationDateYYYYMMDD | 2018-10-01 |
| PublicationDate_xml | – month: 10 year: 2018 text: 2018-10-01 day: 01 |
| PublicationDecade | 2010 |
| PublicationPlace | New York |
| PublicationPlace_xml | – name: New York |
| PublicationTitle | IEEE journal of selected topics in signal processing |
| PublicationTitleAbbrev | JSTSP |
| PublicationYear | 2018 |
| Publisher | IEEE The Institute of Electrical and Electronics Engineers, Inc. (IEEE) |
| Publisher_xml | – name: IEEE – name: The Institute of Electrical and Electronics Engineers, Inc. (IEEE) |
| References | ref57 ref13 ref12 raskutti (ref41) 2009 ref52 ref11 ref54 schölkopf (ref56) 2001 ref17 ref19 tran (ref6) 2016 kallus (ref10) 2017 ref51 alaa (ref15) 2017 ref46 ref45 ref48 ref47 ref42 rasmussen (ref53) 2006; 1 ref44 ref43 bonilla (ref55) 2008 ref49 johansson (ref18) 2016 ref8 bernardo (ref34) 1998; 6 barron (ref22) 1999; 27 ref4 ref3 cover (ref32) 2012 li (ref9) 2017 atan (ref14) 0 ref35 ref37 bottou (ref2) 2013; 14 ref36 van der vaart (ref31) 0; 3 ref33 ref1 ref39 alaa (ref5) 2017 ref38 wager (ref7) 2017 xie (ref20) 2013; 110 pati (ref50) 2015; 16 shalit (ref30) 2017 linero (ref24) 2017 matthews (ref58) 2017; 18 ref26 ref25 vaart (ref28) 2011; 12 hahn (ref16) 2017 ref21 ref27 ref29 rockova (ref23) 2017 vosburg (ref40) 1966; 17 |
| References_xml | – ident: ref36 doi: 10.1073/pnas.1510489113 – ident: ref4 doi: 10.1002/sim.6265 – start-page: 1789 year: 2017 ident: ref10 article-title: Recursive partitioning for personalization using observational data publication-title: Proc Int Conf Mach Learn – year: 2017 ident: ref5 article-title: Bayesian inference of individualized treatment effects using multi-task Gaussian processes publication-title: Adv Neural Inf Process Syst – ident: ref37 doi: 10.1109/18.54897 – volume: 1 year: 2006 ident: ref53 publication-title: Gaussian Processes for Machine Learning – ident: ref1 doi: 10.1002/sim.4322 – start-page: 3076 year: 2017 ident: ref30 article-title: Estimating individual treatment effect: Generalization bounds and algorithms publication-title: Proc 34th Int Conf Mach Learn – year: 0 ident: ref14 article-title: Deep-treat: Learning optimal personalized treatments from observational data using neural networks publication-title: Proc Assoc Adv Artif Intell – volume: 110 start-page: 6262 year: 2013 ident: ref20 article-title: Population heterogeneity and causal inference publication-title: Proc Nat Acad Sci doi: 10.1073/pnas.1303102110 – volume: 16 start-page: 2837 year: 2015 ident: ref50 article-title: Optimal Bayesian estimation in random covariate design with a rescaled Gaussian process prior. publication-title: J Mach Learn Res – year: 2017 ident: ref23 article-title: Posterior concentration for Bayesian regression trees and their ensembles publication-title: arXiv 1708 08734 – ident: ref35 doi: 10.1109/TIT.2017.2773474 – start-page: 3020 year: 2016 ident: ref18 article-title: Learning representations for counterfactual inference publication-title: Proc Int Conf Mach Learn – year: 2017 ident: ref15 article-title: Deep counterfactual networks with propensity-dropout publication-title: Proc ICML Workshop Principled Approaches Deep Learn – volume: 12 start-page: 2095 year: 2011 ident: ref28 article-title: Information rates of nonparametric Gaussian process methods publication-title: J Mach Learn Res – ident: ref49 doi: 10.1214/08-EJS273 – ident: ref46 doi: 10.1006/jmva.1997.1659 – volume: 14 start-page: 3207 year: 2013 ident: ref2 article-title: Counterfactual reasoning and learning systems: The example of computational advertising publication-title: J Mach Learn Res – ident: ref51 doi: 10.1214/15-AOS1341 – ident: ref11 doi: 10.1080/01621459.1984.10478078 – ident: ref8 doi: 10.1198/jcgs.2010.08162 – ident: ref48 doi: 10.1214/074921708000000156 – ident: ref33 doi: 10.1109/TIT.2007.915707 – volume: 27 start-page: 536 year: 1999 ident: ref22 article-title: The consistency of posterior distributions in nonparametric problems publication-title: Ann Statist doi: 10.1214/aos/1018031206 – start-page: 153 year: 2008 ident: ref55 article-title: Multi-task Gaussian process prediction publication-title: Proc Adv Neural Inf Process Syst – year: 2017 ident: ref24 article-title: Bayesian regression tree ensembles that adapt to smoothness and sparsity publication-title: arXiv 1707 09461 – ident: ref54 doi: 10.1561/2200000036 – start-page: 930 year: 2017 ident: ref9 article-title: Matching on balanced nonlinear representations for treatment effects estimation publication-title: Proc Adv Neural Inf Process Syst – start-page: 1563 year: 2009 ident: ref41 article-title: Lower bounds on minimax rates for nonparametric regression with additive sparsity and smoothness publication-title: Proc Adv Neural Inf Process Syst – volume: 6 year: 1998 ident: ref34 article-title: Information-theoretic characterization of bayes performance and the choice of priors in parametric and nonparametric problems publication-title: Bayesian Statistics – ident: ref29 doi: 10.1214/aos/1017939142 – volume: 17 start-page: 665 year: 1966 ident: ref40 article-title: Metric entropy of certain classes of Lipschitz functions publication-title: Proc Amer Math Soc doi: 10.1090/S0002-9939-1966-0193624-0 – ident: ref57 doi: 10.2202/1557-4679.1308 – ident: ref26 doi: 10.1214/aos/1176345969 – ident: ref43 doi: 10.1007/BF00050848 – ident: ref17 doi: 10.1257/aer.p20171038 – year: 2017 ident: ref16 article-title: Bayesian regression tree models for causal inference: Regularization, confounding, and heterogeneous effects doi: 10.2139/ssrn.3048177 – ident: ref12 doi: 10.1037/h0037350 – year: 2016 ident: ref6 article-title: Model criticism for Bayesian causal inference – year: 2012 ident: ref32 publication-title: Elements of Information Theory – ident: ref45 doi: 10.1007/s11749-007-0075-5 – volume: 3 year: 0 ident: ref31 publication-title: Asymptotic Statistics – ident: ref47 doi: 10.1214/009053607000000613 – ident: ref42 doi: 10.1214/009053604000000300 – year: 2017 ident: ref7 article-title: Estimation and inference of heterogeneous treatment effects using random forests publication-title: J Amer Statist Assoc – ident: ref52 doi: 10.1214/11-AOS920 – ident: ref25 doi: 10.1214/15-EJS1078 – ident: ref39 doi: 10.1007/978-94-017-2973-4_7 – ident: ref19 doi: 10.1002/sim.7623 – ident: ref13 doi: 10.3982/ECTA11293 – ident: ref44 doi: 10.1080/10485252.2017.1339309 – volume: 18 start-page: 1299 year: 2017 ident: ref58 article-title: Gpflow: A Gaussian process library using tensorflow publication-title: J Mach Learn Res – ident: ref27 doi: 10.1214/aos/1176345206 – ident: ref3 doi: 10.1177/0081175012452652 – ident: ref21 doi: 10.1080/10618600.2017.1356325 – ident: ref38 doi: 10.1070/SM1967v002n03ABEH002343 – start-page: 416 year: 2001 ident: ref56 article-title: A generalized representer theorem publication-title: Computational Learning Theory doi: 10.1007/3-540-44581-1_27 |
| SSID | ssj0057614 |
| Score | 2.4127903 |
| Snippet | We investigate the problem of estimating the causal effect of a treatment on individual subjects from observational data; this is a central problem in various... |
| SourceID | proquest crossref ieee |
| SourceType | Aggregation Database Enrichment Source Index Database Publisher |
| StartPage | 1031 |
| SubjectTerms | Algorithms Asymptotic methods Bayes methods Bayesian analysis Bayesian nonparametrics Bias causal effect inference Data models Divergence Domains Empirical analysis Estimation Gaussian process Gaussian processes Hilbert space Inference Inference algorithms Information rates Information theory Machine learning Maximization multitask learning Nonparametric statistics Optimization selection bias Signal processing algorithms Smoothness |
| Title | Bayesian Nonparametric Causal Inference: Information Rates and Learning Algorithms |
| URI | https://ieeexplore.ieee.org/document/8387845 https://www.proquest.com/docview/2115837505 |
| Volume | 12 |
| WOSCitedRecordID | wos000446341300016&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| hasFullText | 1 |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| journalDatabaseRights | – providerCode: PRVIEE databaseName: IEEE Xplore customDbUrl: eissn: 1941-0484 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0057614 issn: 1932-4553 databaseCode: RIE dateStart: 20070101 isFulltext: true titleUrlDefault: https://ieeexplore.ieee.org/ providerName: IEEE |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LT8JAEJ4g8aAHX2hE0ezBmxa27bbdekMi0QshqAm3Zp9IAsXwMPHfu1u2RKMx8dbDTrOd2Z1HZ74ZgCssCbd95bwgJNIj1CceCwLhmZukpYqpoFgXwyaSXo8Oh2m_AjcbLIxSqig-U037WOTy5Uys7K-yFg1pQkm0BVtJEq-xWqXWNW6z7zLIgUeiKCwBMjhtmSP-1LdVXLRplLHNLH0zQsVUlR-quLAv3f3_7ewA9pwfidprwR9CReVHsPulu2ANBnfsQ1mMJOrNctvie2qnZwnUYauFIX0soX63yGGSrIzQwDqfiOUSudarI9SejGbz8fJ1ujiGl-79c-fBcyMUPBGk0dJTCdbCpypMfU1izkw4KqQJsrimJvRjOhYy4EzyiOGUYyVZbNyHkEWKc8UjnoYnUM1nuToFpBVJhBAKM0ZJaExa6BNNBY815qkJI-vglzzNhOsvbsdcTLIizsBpVsghs3LInBzqcL2heVt31_hzdc1yfrPSMb0OjVJ0mbuAi8xsKDIfa_y7s9-pzmHHvntdl9eA6nK-UhewLd6X48X8sjhbnz1CzXE |
| linkProvider | IEEE |
| linkToHtml | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LS8NAEB58gXrwVcVq1T1409RNskk33rQoLdZSqkJvYZ9aqKn0Ifjv3U03RVEEbznskM3M7jwy880AnGJJuO0r5wUhkR6hPvFYEAjP3CQtVUwFxTofNlFrt2mvl3QW4HyOhVFK5cVnqmof81y-HIqp_VV2QUNaoyRahGU7OcuhtQq9axxn3-WQA49EUVhAZHByYQ75Q8fWcdGqUcc2t_TNDOVzVX4o49zC3G7-b29bsOE8SXQ1E_02LKhsB9a_9BcsQfeafSiLkkTtYWabfL_a-VkC1dl0bEibBdjvEjlUkpUS6lr3E7FMItd89RldDZ6Ho_7k5XW8C0-3N4_1hueGKHgiSKKJp2pYC5-qMPE1iTkzAamQJszimprgj-lYyIAzySOGE46VZLFxIEIWKc4Vj3gS7sFSNszUPiCtSE0IoTBjlITGqIU-0VTwWGOemECyDH7B01S4DuN20MUgzSMNnKS5HFIrh9TJoQxnc5q3WX-NP1eXLOfnKx3Ty1ApRJe6KzhOzYYi87HGwzv4neoEVhuP96201WzfHcKafc-sSq8CS5PRVB3Binif9Mej4_ycfQIaKtC6 |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Bayesian+Nonparametric+Causal+Inference%3A+Information+Rates+and+Learning+Algorithms&rft.jtitle=IEEE+journal+of+selected+topics+in+signal+processing&rft.au=Alaa%2C+Ahmed+M.&rft.au=van+der+Schaar%2C+Mihaela&rft.date=2018-10-01&rft.pub=IEEE&rft.issn=1932-4553&rft.volume=12&rft.issue=5&rft.spage=1031&rft.epage=1046&rft_id=info:doi/10.1109%2FJSTSP.2018.2848230&rft.externalDocID=8387845 |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1932-4553&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1932-4553&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1932-4553&client=summon |